Column start (1) greater than line length (0)

[bigslycat]

Error: Column start (1) greater than line length (0)
  at Object.rangeFromLineNumber (/home/cat/.atom/packages/linter-eslint/node_modules/atom-linter/lib/index.js:179:11)
  at /home/cat/.atom/packages/linter-eslint/lib/main.js:142:33
  at Array.map (native)
  at /home/cat/.atom/packages/linter-eslint/lib/main.js:125:27

Atom v1.10.1, Ubuntu 16.10

[slackday]

I get the same error on Atom v.1.10.2, Mac OSX 10.10.5 after updating this plugin to v7.3.0 today.

Don't really know where to start looking for source of this error or what other info might be helpful.

[dev-pro-alexander-subota]

Same issue, Ubuntu 14.04, Atom 1.9.7, updated to plugin v7.3.0

[IanVS]

I am not seeing any problems with Atom 1.10.1, mac OSX, and linter-eslint 7.3.0.

For those of you having this problem, can you try restarting Atom and confirm that it still occurs?

If so, some more details would be great. e.g. does this happen on every file you try to open, or only certain files?

[dev-pro-alexander-subota]

@IanVS It happens when user write the code and code have unclosed brackets.

Just updated Atom to 1.10.2, issue is still presents

[rockerest]

Also just updated to 1.10.2 and this issue persists.

For the record, I don't have to type anything, it happens on initial load of the editor, too.
Adding a blank line anywhere causes it. I don't think it has to do with unclosed braces.

[rockerest]

Just restarted to test, and the issue still persists after a reboot.

Running Xubuntu 16.04 (Ubuntu 16.04 base).

[halcarleton]

I'm getting this error even after restarting atom. I had a similar issue with linter-jshint before I switched to linter-eslint. This latest update is the first time I have seen this issue with linter-eslint.

Here's a gif of the easiest way I have found to recreate the issue:

Here is my .eslintrc.json file.

linter-eslint 7.3.0
eslint        3.4.0
Ubuntu        16.04

[Arcanemagus]

Looking into this now.

[Arcanemagus]

@halleycarleton Your specific case is due to no-multiple-empty-lines reporting a column of 2 on blank lines... which don't have that many columns.

@dev-pro-alexander-subota and @rockerest it looks like you are both hitting the same thing as well.

@bigslycat I'm not sure what you are hitting as you provided no information necessesary to reproduce the error (file that shows it + config for eslint + version used), but I'm guessing you are hitting the same problem.

I'll get a nice clean reproduction case and file this over on ESLint.

[IanVS]

@Arcanemagus Perhaps we should at least add a safety check to make sure the column number is valid and avoid throwing an error?

[rockerest]

@Arcanemagus Can (probably) confirm that it's the no-multiple-empty-lines rule, causing other errors doesn't result in the popup, it looks like it's only errors when the line length is 0.

Update for anyone coming here later: It's not just lines where the length is zero, but any rule checking (apparently) at the end of the line. Have seen a bunch of places where it's referencing column 51 when there are only 50 columns, etc.

[Arcanemagus]

Filed over there, turns out this has been broken since eslint@1.8.0, I have no idea why this wasn't blowing up on everyone long ago 😕.

[lmj0011]

reverting back to linter-eslint@7.2.4 is what I'm doing until a patch is implemented

[Arcanemagus]

ESLint v3.5.0 includes the fix by @IanVS for this bug, update the local copy in your projects and you should stop seeing this error 😉.

[sk22]

Nope, still facing the same issue.

> .\node_modules\.bin\eslint --version
v3.5.0

Error: Column start (1) greater than line length (0)
    at Object.rangeFromLineNumber (C:\Users\samuel\.atom\packages\linter-eslint\node_modules\atom-linter\lib\index.js:179:11)
    at C:\Users\samuel\.atom\packages\linter-eslint\lib\main.js:142:33
    at Array.map (native)
    at C:\Users\samuel\.atom\packages\linter-eslint\lib\main.js:125:27

[sk22]

I don't even see eslint's path in the error trace. So why should eslint be responsible for the issue, or is the trace incomplete?

[scholtzm]

ESLint itself provides the invalid marker, e.g. column position 1 for line with 0 characters.

[sk22]

But shouldn't the problem have been fixed with 3.5.0?

[IanVS]

@22sk, would you mind linting the file which is throwing the error using the eslint cli, and post the results? It could well be another rule that is causing your problems. The only one fixed in ESLint so far is no-multiple-empty-lines, but there are likely others as well.

[sk22]

@IanVS The message appears like completely randomly, if any line length changes. The CLI's output looks just as it should. No thrown errors or such.

[Arcanemagus]

@22sk did you restart Atom after updating eslint in your project? It loads eslint per path, and won't update the loaded version until you restart Atom.

[jokeyrhyme]

I've restarted and I'm no longer getting the error related to no-multiple-empty-lines. Yay! Thanks!

I did see it for one other error, but I can't remember how I reproduced it. When I figure it out, I'll report it against upstream ESLint.

[sk22]

@Arcanemagus I did...

[scholtzm]

FWIW, ESLint v3.5.0 fixes the issue on my end.

edit. After using Atom for a while, I'm still getting this error occasionally, but with a different (higher) column numbers than before. I'm not able to reproduce this consistently though.

I guess #689 is the error I'm seeing now.

[Arcanemagus]

v7.3.1 is released which should show the offending error in the message shown to you.

As the bug causing the originally reported issue has been fixed in eslint@3.5.0 I'm locking this issue. Please file a new issue here if you need help tracking down where the problem is, or preferably over on ESLint as you should now be able to tell what the offending rule is 😉.

Thanks everyone for taking the time to help track this down and report it!